Bosco S. Tjan

Viewpoint Dependence in Visual and Haptic Object Recognition

On the whole, people recognize objects best when they see the objects from a familiar view and worse when they see the objects from views that were previously occluded from sight. Unexpectedly, we found haptic object recognition to be viewpoint-specific as well, even though hand movements were unrestricted. This viewpoint dependence was due to the hands preferring the back “view” of the objects. Furthermore, when the sensory modalities (visual vs. haptic) differed between learning an object and recognizing it, recognition performance was best when the objects were rotated back-to-front between learning and recognition. Our data indicate that the visual system recognizes the front view of objects best, whereas the hand recognizes objects best from the back.

Spatial-frequency characteristics of letter identification in central and peripheral vision

Spatial-frequency characteristics of letter identification are much better understood in the fovea than in the periphery. The purpose of this study was to compare the spatial-frequency characteristics of letter identification in central and peripheral vision. We measured contrast thresholds for identifying single, Times-Roman lower-case letters that were spatially band-pass filtered. Each of the 26 letters was digitally filtered with a set of nine cosine log filters, with peak object spatial frequencies ranging from 0.63 to 10 c/letter, in half-octave steps. Bandwidth of the filters was 1 octave. Three observers with normal vision were each tested monocularly at the fovea, and at 5 degrees and 10 degrees in the inferior visual field. Letter sizes were 0.2, 0.4 and 0.6 log units larger than high contrast, unfiltered acuity letters. Plots of contrast sensitivity for letter identification vs. frequency of the band-pass filters exhibit spatial tuning. In general, the spatial-frequency characteristics of letter identification are fundamentally identical between central and peripheral vision. These characteristics include the scaling of the peak frequency of the spatial-tuning functions with letter size and the bandwidth of the tuning functions. The only difference between the fovea and the periphery is that for the same physical letter size, peak sensitivity of the spatial-tuning functions occurs at a higher retinal frequency at the fovea than in the periphery. To test whether or not the contrast sensitivity function (CSF) can account for the differences in the spatial-frequency characteristics of letter identification between central and peripheral vision, we incorporated a human CSF into an ideal-observer model, and tested the performance of this ideal-observer on the same letter identification task used with the human observers. Data from this CSF-ideal-observer resemble closely those of human observers, suggesting that the spatial-frequency characteristics of human letter identification can be accounted for by the CSF and the letter-identity information, without invoking selection among narrow-band spatial-frequency channels.

Mr. Chips 2002: new insights from an ideal-observer model of reading.

The integration of visual, lexical, and oculomotor information is a critical part of reading. Mr. Chips is an ideal-observer model that combines these sources of information optimally to read simple texts in the minimum number of saccades. This model provides a computational framework for interpreting human reading saccades in both normal and low vision. The purpose of this paper is to report performance of the model for conditions emulating reading with normal vision--a visual span of nine characters, multiplicative saccade noise with a standard deviation of 30%, and texts based on three full-length childrens books. Comparison of fixation locations by humans and Mr. Chips revealed: (1) that both exhibit very similar word-skipping behavior; (2) both show initial fixations near the center of words, but with a systematic difference suggestive of an asymmetry in the human visual span; and (3) differences in the pattern of refixations within words that may uncover non-optimal lexical inference by human readers. A human context effect--30% difference in mean saccade size between continuous text and random sequences of words--was very similar to the 25% effect for the model associated with a corresponding difference in the predictability of text words. Overall, our findings show that many of the complicated aspects of human reading saccades can be explained concisely by early information-processing constraints.

Blood oxygenation level-dependent contrast response functions identify mechanisms of covert attention in early visual areas

Covert attention can lead to improved performance in perceptual tasks. The neural and functional mechanisms of covert attention are still under investigation. Using both rapid event-related and mixed designs, we measured the blood oxygenation level-dependent functional MRI contrast response functions over the full range of contrast (0–100%) in the retinotopically defined early visual areas (V1, V2, V3, V3A, and V4) in humans. Covert attention increased both the baseline activities and contrast gains in the five cortical areas. The effect on baseline can be decomposed into a transient trial-by-trial component and a component across an entire attention block. On average, increase in contrast gain accounted for ≈88.0%, 28.5%, 12.7%, 35.9%, and 25.2% of the trial-by-trial effects of attention in the five areas, respectively, and 22.2%, 12.8%, 7.4%, 19.7%, and 17.3% of the total effects of attention in those areas, consistent with single-unit findings in V4 and MT. The results provide strong evidence for a stimulus enhancement mechanism of attention as demonstrated in various behavioral studies.

What makes faces special

What may be special about faces, compared to non-face objects, is that their neural representation may be fundamentally spatial, e.g., Gabor-like. Subjects matched a sequence of two filtered images, each containing every other combination of spatial frequency and orientation, of faces or non-face 3D blobs, judging whether the person or blob was the same or different. On a match trial, the images were either identical or complementary (containing the remaining spatial frequency and orientation content). Relative to an identical pair of images, a complementary pair of faces, but not blobs, reduced matching accuracy and released fMRI adaptation in the fusiform face area.

Saccade-confounded image statistics explain visual crowding

Processing of shape information in human peripheral visual fields is impeded beyond what can be expected by poor spatial resolution. Visual crowding, the inability to identify objects in clutter, has been shown to be the primary factor limiting shape perception in peripheral vision. Despite the well-documented effects of crowding, its underlying causes remain poorly understood. Given that spatial attention both facilitates learning of image statistics and directs saccadic eye movements, we propose that the acquisition of image statistics in peripheral visual fields is confounded by eye-movement artifacts. Specifically, the image statistics acquired under a peripherally deployed spotlight of attention are systematically biased by saccade-induced image displacements. These erroneously represented image statistics lead to inappropriate contextual interactions in the periphery and cause crowding.

The Perception of a Face Is No More Than the Sum of Its Parts

When you see a person’s face, how do you go about combining his or her facial features to make a decision about who that person is? Most current theories of face perception assert that the ability to recognize a human face is not simply the result of an independent analysis of individual features, but instead involves a holistic coding of the relationships among features. This coding is thought to enhance people’s ability to recognize a face beyond what would be expected if each feature were shown in isolation. In the study reported here, we explicitly tested this idea by comparing human performance on facial-feature integration with that of an optimal Bayesian integrator. Contrary to the predictions of most current notions of face perception, our findings showed that human observers integrate facial features in a manner that is no better than would be predicted by their ability to use each individual feature when shown in isolation. That is, a face is perceived no better than the sum of its individual parts.

Digital Sign System for Indoor Wayfinding for the Visually Impaired

Mobility challenges and independent travel are major concerns for blind and visually impaired pedestrians [1][2]. Navigation and wayfinding in unfamiliar indoor environments are particularly challenging because blind pedestrians do not have ready access to building maps, signs and other orienting devices. The development of assistive technologies to aid wayfinding is hampered by the lack of a reliable and costefficient method for providing location information in an indoor environment. Here we describe the design and implementation of a digital sign system based on low-cost passive retro-reflective tags printed with specially designed patterns that can be readily detected and identified by a handheld camera and machine-vision system. Performance of the prototype showed the tag detection/recognition system could cope with the real-world environment of a typical building.

Human efficiency for recognizing and detecting low-pass filtered objects

Recently, Tjan, Braje, Legge and Kersten [(1995) Vision Research, 35, 3053-3069] found that human efficiency for object recognition was less than 10%, indicating that humans fail to use much of the information available to an ideal observer. We examine two explanations for these low efficiencies: (1) humans are inefficient in using high spatial-frequency information; and (2) humans are inefficient in detecting image samples. We tested the first possibility by measuring human efficiency for recognizing low-pass filtered objects, rendered as line drawings and silhouettes, in luminance noise. Efficiency did not improve when high frequencies were removed, and the first explanation was rejected. We tested the second explanation by comparing efficiencies for object detection and recognition. Recognition efficiency was higher than detection efficiency for silhouettes but not line drawings, showing that detection efficiency does not place a ceiling on recognition efficiency. The results indicate that human vision is designed to extract image features, such as contours, that enhance recognition. A computer simulation suggests that this can occur if the observer views the world through a band-pass spatial-frequency channel.

Rapid and Persistent Adaptability of Human Oculomotor Control in Response to Simulated Central Vision Loss

The central region of the human retina, the fovea, provides high-acuity vision. The oculomotor system continually brings targets of interest into the fovea via ballistic eye movements (saccades). Thus, the fovea serves both as the locus for fixations and as the oculomotor reference for saccades. This highly automated process of foveation is functionally critical to vision and is observed from infancy. How would the oculomotor system adjust to a loss of foveal vision (central scotoma)? Clinical observations of patients with central vision loss suggest a lengthy adjustment period, but the nature and dynamics of this adjustment remain unclear. Here, we demonstrate that the oculomotor system can spontaneously and rapidly adopt a peripheral locus for fixation and can rereference saccades to this locus in normally sighted individuals whose central vision is blocked by an artificial scotoma. Once developed, the fixation locus is retained over weeks in the absence of the simulated scotoma. Our data reveal a basic guiding principle of the oculomotor system that prefers control simplicity over optimality. We demonstrate the importance of a visible scotoma on the speed of the adjustment and suggest a possible rehabilitation regimen for patients with central vision loss.